Hi

We have very serious situation in client production system.
After one year of proper system working, one day, on VPMS there starting to appear following errors:
error.badfetch.http.905

We try to investigate network problems by making ping call from MPP to application server (Ibm WebSphere) but we didn't find any problems.

We also try to avoid using https an LoadBalancer by calling appliaction directly - that also doesn't work.

We changed AVP fetch timeout to 45 seconds , and that allows us to avoid error.badfetch.http.905 problems - but application reponse time was verry slowly. As we looked into applicaiton logs, we see that time interval beetwen each following logs was very long (about one second) even if there was no communication to backend, so it seems that applciocation with now reason slow down.

We try to invoke appliaction by calling from standard web browser, and application respond very quick as it should.


What are differences beetwen VP invokation and normal application invokation. We are aware, that from webrowser DD turn in somtehing like debug mode and avoids using CTI Connector. We have checked connection beewten CTIconnector and AES and we haven't seen any problems.
Are there any differences wchich could cause such slow down ?

Did something change codewise or on the appserver or AVP when the problems started? Any backend servers involved? Webapps, database ops, etc?

Have you reset the app server and AVP servers since these problems started up?

It's rather hard to diagnose a problem like this without alot more information. There are no known issues related to this.

I would start by comparing the Voice Browser logs when it performs the page fetch and how quickly the app is getting those requests and how long it takes to respond. See if you can narrow down any particular requests that are slow. If all seem slow, determine where the slowdown is occuring? It could be network related.

The code wasn't changed. Application slowdown occurs very before any backend operation are performed.

We restarted Application Server, MPP, VPMS, AES and it didn't help at all.

We sniffed network traffic bewteen MPP and WAS. HTTP requests where transported immediately, so we exluded network related problems.

We know, that before malfunction, there was DNS server restart. Can it have any fluance on Dialog Designer runtime ? Maybe some DD component (CTIConnector) would timeouts on DNS resolv requests ?

Can you pinpoint in your application when you see the slowdown? What is the app doing at the time?

Hello,
I am experience exact same issue with another customer.
The error seen in MPP logs is different though:
error.badfetch.http.404
Its happening intermittently at different times.

The malfunction where lasting for 16 hours. After that, application started to work itself as it should.

Now we are tryging to find root cause of the slowdown to eliminate it in the future. We aren't able to reproduce the problem now.

Michal - Ok. Not sure we have much to go by as far as debugging the problem. If you are trying to reproduce, I recommend getting app log going and capture browser logs too if the problem crops up again.

Ankita - Please start a new thread to avoid confusion. You will need to gather more logs (browser and app) to determine why the bad fetch might be occuring.

Hi Neil,

I have found, that at the crush time there was following ERROR loged in CTIConnector.log

CTICallOperation.verifyEncryptedValue:exceeded: 57441

Can it have any influance on other components ?